Erasca, a biotech startup run by the former CEO of Ignyta, debuted in December with $42 million and an audacious plan to “erase cancer,” but provided little explanation as to how.
Erasca is still keeping many of those details under wraps. But Jonathan Lim, the company’s co-founder and executive chairman, this week outlined, at least in broad strokes, Erasca’s strategy: a two-pronged approach, involving drug making and big data analytics.
That plan folds the startup into the ranks of biopharma firms that are trying to use artificial intelligence, machine learning, and other computing tools to discover drugs—specifically, in Erasca’s case, precision cancer therapies—faster than with standard approaches. The company will use those tools to supplement its research and development and advance drugs it develops in either in-house or acquired through deals, Lim says.
Now the company, which began operations in October, has $22 million more in its bank account to help with the wide-ranging effort. The cash comes from new investors, including Chicago-based Arch Venture Partners, Silicon Valley’s Andreessen Horowitz, and New York firm Reneo Capital. Other new private and strategic investors, which Erasca didn’t name, also participated in the round, announced this week. City Hill Ventures, an investment firm started by Lim, and Cormorant Asset Management led Erasca’s initial $42 million haul, which was announced in December.
In 2016, the Tufts Center for the Study of Drug Development estimated that it takes often more than a decade and about $2.6 billion to bring a new drug to market. Companies have been searching for years for ways to speed up and improve the success rates of R&D, and recently have begun zeroing in on machine learning and AI.
As this Nature article notes, these efforts have much to prove, however. Excitement over computer-aided drug design began some 30 years ago but has yet to ameliorate a persistent slowdown in R&D productivity.
Still, a slew of companies, large and small, are now looking to take advantage of recent advances in computing power and data storage in a hunt for new treatments.
Summit, NJ-based Celgene (NASDAQ: CELG), for example, struck a multi-year agreement with Cambridge, MA-based GNS Healthcare in late 2016 to license the firm’s “Reverse Engineering and Forward Simulation” software, which uses machine learning technology. Roche’s Genentech unit has been working with GNS since 2017 to try to discover cancer drugs. Shortly after the Celgene-GNS announcement, Pfizer (NYSE: PFE) said it would use Watson, the computer system developed by Armonk, NY-based IBM (NYSE: IBM), to accelerate its efforts to develop cancer immunotherapies.
Others with in-house AI drug discovery tools are in the mix too, among them San Francisco’s Engine Biosciences, BenevolentAI of London, and Boston-based Berg, a biopharma company backed by real estate billionaire Carl Berg.
Now comes Erasca, which is using a computing system called Oncology Pattern Recognition Algorithm, or OPRA. Lim credits Erasca’s vice president of biology, Robert Shoemaker, with its development. Shoemaker worked under Lim at Ignyta, heading its computational biology group. While at the company, he played a key role in the development of entrectinib, the targeted cancer drug that led Roche to buy the firm for $1.7 billion in 2017. Shoemaker strategized which types of patients to treat with entrectinib, built a companion diagnostic test for the drug, and more, according to Erasca.
Lim wouldn’t detail exactly how OPRA works or what would give it an edge over other AI-based approaches.
“I want to be careful not to provide too much specificity at this point because …we have the platform, but we’re also building it out,” Lim said.
But he did say the system helps comb through publicly available and proprietary information on cancer genomics, cancer drug targets, and biological pathways of interest. The hope is with those insights, and more broadly, machine learning and automation, Erasca can pinpoint “what parts of discovery we can just fundamentally change or accelerate.”
Ideally, OPRA would “augment and complement” the work of Erasca’s team and help reveal “uncommon insights [into cancer biology] that we would have not found” otherwise, he says.
Human trials at the company are at least a couple years out. With the latest cash infusion, there’s enough funding to carry Erasca through late 2022, Lim says.
In the meantime, Lim notes that Arch and a16z are involved with several other startups—such as AI expert Daphne Koller’s Insitro—that also aim to use advanced computing tools in drug discovery, which should help the company with its own efforts.
“I think we can leverage some of the best practices across their portfolio companies as well to achieve a better answer for patients with cancer,” Lim says.
A16z invested in Erasca out of its second bio fund, a $450 million investment fund launched in 2017. (Its first, started in 2015, totaled $200 million.) Other companies in the a16z bio portfolio, in addition to Insitro, include Boston-based synthetic biology startup Asimov and CAMP4, a startup in Cambridge, MA, that’s developed software it says can help researchers understand how genes are regulated.
Jorge Conde, a general partner at a16z, says he believes in Erasca because its founding team has a track record of finding new drugs, and it is built around technologies that could help broaden our understanding of cancer biology. AI might help unearth new ways to get at known cancer drivers, or find new important targets that scientists haven’t yet discovered, he says.
“If there’s one thing AI can do well, it is to help make sense of complex data sets that may be beyond human grasp,” Conde says.